ComfyUI

ComfyUI is a local visual AI workspace for artists and technical teams who want to control how images, video, audio, 3D models and text are made. Its node canvas shows each model and processing step, so users can build and adjust workflows without writing code. It runs on your hardware.

The local software works on Windows, macOS and Linux and is open source under GPL-3.0. Its core can run fully offline. Optional API nodes connect to outside model services, while Comfy Cloud offers a hosted browser option. A separate developer platform lets teams deploy workflows as production APIs.

The graph supports reusable workflows for generation and editing. Community templates offer a starting point, and App Mode presents a simpler view of a finished workflow while keeping the underlying nodes available. Local tools include inpainting, masks, compositing, upscaling and frame interpolation. Workflows can be saved as JSON, and custom nodes extend what the canvas can do.

Native model support includes Stable Diffusion, Flux, Qwen Image, Wan, LTX-Video, Hunyuan3D and Gemma. Partner nodes provide access to closed models such as Seedance and Nano Banana. ComfyUI supports NVIDIA, AMD and Intel GPUs through its available builds and backends. Model offloading and memory management help it run larger open models on available hardware; requirements vary by workflow and model.